Oracle Database In-Memory: In-Memory Aggregation (White Paper)

Oracle White Paper
Oracle Database In-Memory: In-Memory Aggregation

Introduction
In-Memory Aggregation, a feature of the Oracle Database In-Memory, is the new best practice for executing star queries in the Oracle Database. In-Memory Aggregation provides new SQL execution operations that accelerate the performance of wide range of analytic queries against star and similar schemas. In-Memory Aggregation optimizes joins between dimension and fact tables, and computes aggregations using CPU efficient algorithms. The relative advantage of In-Memory Aggregation increases as queries become more difficult, joining more dimension tables to the fact table and aggregating over more grouping columns and more fact rows. In-Memory Aggregation provides the greatest benefit when used with together with tables populated in the in-memory column store, though In-Memory 

Aggregation also delivers big benefits when used with conventional tables stored on disk.
In-Memory Aggregation will be automatically chosen by the SQL optimizer based on execution cost.
Applications do not need to be modified to use In-Memory Aggregation.

http://www.oracle.com/technetwork/database/bi-datawarehousing/inmemory-aggregation-twp-01282015-2412192.pdf